Repository navigation
fastfetch: add new package - #27949
betonmischer86 wants to merge 1 commit into
Conversation
GeorgeSapkin
left a comment
GeorgeSapkin
left a comment
There was a problem hiding this comment.
As with the other PRs, you need to set yourself as the maintainer and don't need to include the license file.
1d8f9ff to
b5ec378
Compare
Done. |
|
It seems that this PR is not merged for some time. Is there is any reason for it? Because tools like |
|
Yes, it is described by this comment #29564 (comment) |
cf6086a to
3ff77a1
Compare
|
@GeorgeSapkin I updated this PR following the recent discussion at #29564. The binary has been trimmed down to 773K, extra dependencies and shell completion files removed. |
|
In the meantime, it's possible to install it from Alpine: #!/bin/ash
apk --allow-untrusted --arch armv7 --no-cache --repository https://dl-cdn.alpinelinux.org/alpine/edge/community fetch fastfetch
apk --allow-untrusted --arch armv7 --no-cache --repository https://dl-cdn.alpinelinux.org/alpine/edge/main fetch hwdata-pci
apk --allow-untrusted add hwdata-pci-*.apk
gzip -dc fastfetch-*.apk | tar -xf - -C /
rm fastfetch-*.apk hwdata-pci-*.apk
fastfetchOS: OpenWrt 25.12.4 armv7l
Host: Linksys WRT3200ACM
Kernel: Linux 6.12.87
Uptime: 2 hours, 6 mins
Packages: 194 (apk)
Shell: ash 1.37.0
Terminal: dropbear
CPU: armada380 (2)
Memory: 258.85 MiB / 495.00 MiB (52%)
Swap: Disabled
Disk (/): 22.92 MiB / 54.65 MiB (42%) - overlay
Locale: C |
|
|
3ff77a1 to
8901ac3
Compare
@CarterLi Thanks, that allowed to shave off an extra 170K. As for the supported package managers, I see no harm in keeping the default cmake options since that doesn't take any extra space. |
openwrt-ai
left a comment
openwrt-ai
left a comment
There was a problem hiding this comment.
Reviewed 1 new commit (new package fastfetch). CI is green across all target architectures and the upstream tag 2.64.2 / PKG_HASH and source URL check out. One inline comment on the CMAKE_OPTIONS flag handling; the commit message matches the diff.
Generated by Claude Code
f217682 to
23341ab
Compare
openwrt-ai
left a comment
openwrt-ai
left a comment
There was a problem hiding this comment.
Reviewed 1 new commit. The only change since my last review is the 2.64.2 → 2.67.0 bump; tag 2.67.0 exists upstream and the rest of the Makefile is unchanged, and the commit message still matches the diff.
Checked against the 2.67.0 tree: the find src/logo/ascii/[a-z]/ prune in Build/Prepare still matches upstream's layout (logo entries are #ifdef-guarded and CMake globs [a-z_]/*.txt), every MODULE_DISABLE_* name still resolves to a directory under src/modules, and elfutils is still needed for ENABLE_ELF. Two of the -DENABLE_* flags went stale with the bump — inline comment. Worth noting that 2.67.0 raises the minimum compiler to GCC 13 / Clang 16 for C23; OpenWrt master ships 13.4 and 15.3, so that is covered, but it rules out older stable toolchains. My earlier comment on -DCMAKE_C_FLAGS / -DCMAKE_EXE_LINKER_FLAGS inside CMAKE_OPTIONS still applies.
Generated by Claude Code
Fastfetch is a Neofetch-like tool for fetching system information and displaying it in a visually-appealing way. Signed-off-by: Valeriy Kosikhin <vkosikhin@gmail.com>
23341ab to
88b27ee
Compare
openwrt-ai
left a comment
openwrt-ai
left a comment
There was a problem hiding this comment.
Reviewed 1 new commit — the head was amended since my last review: 2.67.0 → 2.67.1 plus fixes for both earlier findings.
Both are addressed correctly. -fno-unwind-tables -fno-asynchronous-unwind-tables and -Wl,--strip-all now go through TARGET_CFLAGS / TARGET_LDFLAGS, which include/cmake.mk appends to (CFLAGS= env at cmake.mk#L92, -DCMAKE_EXE_LINKER_FLAGS at #L109) instead of replacing, so the target flags survive now. And ENABLE_VA / ENABLE_DRM_AMDGPU are gone in favour of ENABLE_VADRM / ENABLE_VAX11, both of which exist in 2.67.1 and default to ON on Linux — checked against the tag, where the two old names no longer appear at all.
Tag 2.67.1 exists upstream (released 2026-08-14; the delta over 2.67.0 is bugfixes plus a CMake-time C23 compiler check). PKG_RELEASE:=1 is right for a not-yet-merged new package, and the commit message still matches the diff, so no commit-check findings. I was not able to fetch the tarball in this environment, so PKG_HASH is unverified from my side.
Two optional-dependency questions inline; neither blocks a merge.
Generated by Claude Code
| PKG_LICENSE_FILES:=LICENSE | ||
| PKG_CPE_ID:=cpe:/a:fastfetch-cli:fastfetch | ||
|
|
||
| PKG_BUILD_DEPENDS:=elfutils |
There was a problem hiding this comment.
ENABLE_ELF is left at its upstream default ON — cmake_dependent_option(ENABLE_ELF "Enable libelf" ON "LINUX OR ANDROID OR DragonFly OR Haiku OR GNU" OFF) in fastfetch 2.67.1 CMakeLists.txt — which is why elfutils is needed here. But the package declares no DEPENDS, so libelf is not on the device after install.
If libelf were linked directly, packaging would fail: CheckDependencies in include/package-pack.mk:271-290 rejects any NEEDED library not covered by a declared dependency, and the build is clean. So fastfetch is dlopen-ing libelf at runtime, meaning the ELF code path is compiled in but can never activate on a stock install.
Given how aggressively the rest of this Makefile trims features for size, is keeping it intentional? Either -DENABLE_ELF=OFF (and then PKG_BUILD_DEPENDS:=elfutils can go too) or DEPENDS:=+libelf would make the two ends agree.
Generated by Claude Code
| -DENABLE_ZLIB=OFF \ | ||
| -DENABLE_QUICKJS=OFF \ | ||
| -DENABLE_LIBZFS=OFF \ | ||
| -DENABLE_WCWIDTH=OFF \ |
There was a problem hiding this comment.
nit: ENABLE_LUA is the one optional-library switch left at its upstream default — option(ENABLE_LUA "Enable lua scripting for format strings" ON) in fastfetch 2.67.1 CMakeLists.txt — while every other optional library in this block is explicitly turned off.
That leaves the result dependent on what happens to be in staging: OpenWrt core's lua package installs lua.pc into the staging pkgconfig dir, so a buildroot that has already built lua produces a different (larger) fastfetch than one that hasn't. Deliberate, or should it get -DENABLE_LUA=OFF alongside the rest?
Generated by Claude Code
|
Not sure if this would fulfill your need, but I created an alternative in Openwrt's ucode called owrtfetch. |
8 participants
📦 Package Details
Maintainer: @betonmischer86
Description:
Fastfetch is a Neofetch-like tool for fetching system information and displaying it in a visually-appealing way.
🧪 Run Testing Details
✅ Formalities
If your PR contains a patch:
git am(e.g., subject line, commit description, etc.)
We must try to upstream patches to reduce maintenance burden.